Training in progress, step 10, checkpoint

last-checkpoint/adapter_config.json

@@ -20,13 +20,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
     "k_proj",
-    "q_proj",
-    "up_proj",
     "gate_proj",
+    "o_proj",
     "down_proj",
-    "o_proj"
+    "v_proj",
+    "up_proj",
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8d1651f3601378af486d8705149df9a51b2def80ca66d31c03b802020e5a8db9
+oid sha256:21cb1efe47a7b9ec7582f9c4cad57e9fdc96cddc43b5fe446da11199b86121df
 size 83115256

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0b50128d3b0c6a2edc34f6c7aabbd7e35a705b84c33a17a46e29421d95323100
+oid sha256:ab11797a5a48c8b5091a20e4ab8445d69bbcf0b52c3045370731dea57405fdcc
 size 166351098

last-checkpoint/trainer_state.json

@@ -10,7 +10,7 @@
   "log_history": [
     {
       "epoch": 0.0005925925925925926,
-      "grad_norm": 10.913461685180664,
+      "grad_norm": 10.896183013916016,
       "learning_rate": 5e-07,
       "logits/chosen": -4.444676399230957,
       "logits/rejected": -4.0909342765808105,
@@ -25,7 +25,7 @@
     },
     {
       "epoch": 0.0011851851851851852,
-      "grad_norm": 7.564138412475586,
+      "grad_norm": 7.562353610992432,
       "learning_rate": 4.849231551964771e-07,
       "logits/chosen": -4.530362606048584,
       "logits/rejected": -3.981240749359131,
@@ -40,122 +40,122 @@
     },
     {
       "epoch": 0.0017777777777777779,
-      "grad_norm": 14.393203735351562,
+      "grad_norm": 13.934762001037598,
       "learning_rate": 4.415111107797445e-07,
-      "logits/chosen": -4.539660930633545,
-      "logits/rejected": -4.32346248626709,
-      "logps/chosen": -200.69912719726562,
-      "logps/rejected": -224.77163696289062,
-      "loss": 0.6526,
+      "logits/chosen": -4.541503429412842,
+      "logits/rejected": -4.324827194213867,
+      "logps/chosen": -200.9736328125,
+      "logps/rejected": -224.6981658935547,
+      "loss": 0.6691,
       "rewards/accuracies": 1.0,
-      "rewards/chosen": 0.04995880275964737,
-      "rewards/margins": 0.08375511318445206,
-      "rewards/rejected": -0.03379631042480469,
+      "rewards/chosen": 0.022507095709443092,
+      "rewards/margins": 0.04895630106329918,
+      "rewards/rejected": -0.026449203491210938,
       "step": 3
     },
     {
       "epoch": 0.0023703703703703703,
-      "grad_norm": 12.62684440612793,
+      "grad_norm": 13.249699592590332,
       "learning_rate": 3.75e-07,
-      "logits/chosen": -3.9760360717773438,
-      "logits/rejected": -4.4724321365356445,
-      "logps/chosen": -239.9991455078125,
-      "logps/rejected": -202.63072204589844,
-      "loss": 0.7198,
+      "logits/chosen": -3.9642434120178223,
+      "logits/rejected": -4.461881160736084,
+      "logps/chosen": -239.5960693359375,
+      "logps/rejected": -202.10887145996094,
+      "loss": 0.7259,
       "rewards/accuracies": 0.0,
-      "rewards/chosen": -0.041876986622810364,
-      "rewards/margins": -0.052445217967033386,
-      "rewards/rejected": 0.01056823693215847,
+      "rewards/chosen": -0.001570509746670723,
+      "rewards/margins": -0.06432266533374786,
+      "rewards/rejected": 0.06275215744972229,
       "step": 4
     },
     {
       "epoch": 0.002962962962962963,
-      "grad_norm": 6.745312213897705,
+      "grad_norm": 6.773025989532471,
       "learning_rate": 2.934120444167326e-07,
-      "logits/chosen": -4.136646270751953,
-      "logits/rejected": -4.701557159423828,
-      "logps/chosen": -169.00192260742188,
-      "logps/rejected": -149.2642059326172,
-      "loss": 0.6818,
-      "rewards/accuracies": 0.5,
-      "rewards/chosen": 0.02975158765912056,
-      "rewards/margins": 0.023634720593690872,
-      "rewards/rejected": 0.006116867531090975,
+      "logits/chosen": -4.136728286743164,
+      "logits/rejected": -4.700671672821045,
+      "logps/chosen": -169.1282958984375,
+      "logps/rejected": -149.28570556640625,
+      "loss": 0.6866,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.017114639282226562,
+      "rewards/margins": 0.013147544115781784,
+      "rewards/rejected": 0.003967094700783491,
       "step": 5
     },
     {
       "epoch": 0.0035555555555555557,
-      "grad_norm": 13.741002082824707,
+      "grad_norm": 13.418476104736328,
       "learning_rate": 2.065879555832674e-07,
-      "logits/chosen": -5.057063579559326,
-      "logits/rejected": -5.147495746612549,
-      "logps/chosen": -193.6820526123047,
-      "logps/rejected": -202.20425415039062,
-      "loss": 0.6913,
+      "logits/chosen": -5.0562872886657715,
+      "logits/rejected": -5.145793914794922,
+      "logps/chosen": -194.10568237304688,
+      "logps/rejected": -202.68751525878906,
+      "loss": 0.6884,
       "rewards/accuracies": 0.75,
-      "rewards/chosen": 0.0083160400390625,
-      "rewards/margins": 0.004281995818018913,
-      "rewards/rejected": 0.0040340423583984375,
+      "rewards/chosen": -0.034047700464725494,
+      "rewards/margins": 0.01024474948644638,
+      "rewards/rejected": -0.044292449951171875,
       "step": 6
     },
     {
       "epoch": 0.004148148148148148,
-      "grad_norm": 6.920108795166016,
+      "grad_norm": 6.973136901855469,
       "learning_rate": 1.2500000000000005e-07,
-      "logits/chosen": -4.2448506355285645,
-      "logits/rejected": -3.682173490524292,
-      "logps/chosen": -176.11734008789062,
-      "logps/rejected": -222.982666015625,
-      "loss": 0.7008,
-      "rewards/accuracies": 0.0,
-      "rewards/chosen": 0.028499603271484375,
-      "rewards/margins": -0.015093998052179813,
-      "rewards/rejected": 0.043593600392341614,
+      "logits/chosen": -4.244935035705566,
+      "logits/rejected": -3.681328058242798,
+      "logps/chosen": -176.23780822753906,
+      "logps/rejected": -222.976318359375,
+      "loss": 0.7073,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.016452789306640625,
+      "rewards/margins": -0.027773665264248848,
+      "rewards/rejected": 0.044226452708244324,
       "step": 7
     },
     {
       "epoch": 0.004740740740740741,
-      "grad_norm": 10.856575965881348,
+      "grad_norm": 10.38602066040039,
       "learning_rate": 5.848888922025552e-08,
-      "logits/chosen": -4.413946151733398,
-      "logits/rejected": -4.419940948486328,
-      "logps/chosen": -181.99610900878906,
-      "logps/rejected": -196.20811462402344,
-      "loss": 0.6967,
-      "rewards/accuracies": 0.25,
-      "rewards/chosen": -0.02880115620791912,
-      "rewards/margins": -0.007002831436693668,
-      "rewards/rejected": -0.021798323839902878,
+      "logits/chosen": -4.409693241119385,
+      "logits/rejected": -4.41237735748291,
+      "logps/chosen": -181.64944458007812,
+      "logps/rejected": -196.00564575195312,
+      "loss": 0.6896,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.005864143371582031,
+      "rewards/margins": 0.007416536100208759,
+      "rewards/rejected": -0.001552392728626728,
       "step": 8
     },
     {
       "epoch": 0.005333333333333333,
-      "grad_norm": 9.559263229370117,
+      "grad_norm": 9.169404983520508,
       "learning_rate": 1.507684480352292e-08,
-      "logits/chosen": -4.909343242645264,
-      "logits/rejected": -4.601991176605225,
-      "logps/chosen": -154.6862030029297,
-      "logps/rejected": -179.27401733398438,
-      "loss": 0.6929,
-      "rewards/accuracies": 0.25,
-      "rewards/chosen": 0.0032157916575670242,
-      "rewards/margins": 0.0011837054044008255,
-      "rewards/rejected": 0.0020320871844887733,
+      "logits/chosen": -4.903069972991943,
+      "logits/rejected": -4.5977277755737305,
+      "logps/chosen": -154.88751220703125,
+      "logps/rejected": -179.42715454101562,
+      "loss": 0.695,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.016914749518036842,
+      "rewards/margins": -0.003633500775322318,
+      "rewards/rejected": -0.013281249441206455,
       "step": 9
     },
     {
       "epoch": 0.005925925925925926,
-      "grad_norm": 10.816996574401855,
+      "grad_norm": 10.337571144104004,
       "learning_rate": 0.0,
-      "logits/chosen": -3.8037424087524414,
-      "logits/rejected": -3.683922529220581,
-      "logps/chosen": -232.978515625,
-      "logps/rejected": -254.24078369140625,
-      "loss": 0.7146,
+      "logits/chosen": -3.817504644393921,
+      "logits/rejected": -3.697187900543213,
+      "logps/chosen": -233.053466796875,
+      "logps/rejected": -254.43130493164062,
+      "loss": 0.7085,
       "rewards/accuracies": 0.5,
-      "rewards/chosen": -0.05405044183135033,
-      "rewards/margins": -0.040559008717536926,
-      "rewards/rejected": -0.013491439633071423,
+      "rewards/chosen": -0.06154556944966316,
+      "rewards/margins": -0.029002761468291283,
+      "rewards/rejected": -0.03254280239343643,
       "step": 10
     }
   ],

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:500a27bb31e08eab12d9fb299fbdfd1af99d44a30059f736748d9f929e4cbfa9
+oid sha256:265f72a5d7c9bd4d95478bba01ec7511ef89dd45479699ac483e45fdedd76cef
 size 5944